In-service Radar Fault Prediction Technology Based On Big Data Analysis

With the development of electronic warfare and information warfare technology,the role of military radar in defense is becoming more and more important.In order to ensure the stable,reliable and efficient operation of radar,it is of great significance to study the radar fault prediction technology.The traditional fault prediction technique is mainly based on the empirical model,which can not be influenced by many factors.The prediction accuracy is poor and the false alarm rate is high.With the rapid development of Internet of things and big data technology,the fault prediction technology based on big data has been studied and applied in various fields,which has achieved remarkable results.This article aims at the characteristics of a certain type of active military radar with complex structure,diverse types of faults,diverse working environments,and difficult maintenance.Big data technology is used to analyze and mine radar operating data,usage data,maintenance and maintenance data,fault data,and environmental data.Finding the intrinsic correlation between radar faults and parameters,and designing an algorithm to predict radar faults.Research work has important theoretical and practical significance for safeguarding radar operational performance.The main work and innovation of the thesis is reflected in:1.Taking a certain type of active air defense fire control radar as the research object,it elaborates its working principle,working methods and main structure.Taking its optoelectronic tracking subsystem as an example,the causes of common fault types were analyzed,and the corresponding relationship between the fault and status parameters was determined.2.Multi-dimensional data such as radar real-time operating data,historical fault data,routine maintenance data,and geo-environmental weather data were selected for big data processing and analysis.Firstly,the wavelet threshold denoising method is used to denoise the data,then the Relief-F algorithm is used to characterize the high-level data,and the key influencing factors of the failure of the radar photoelectric tracking subsystem are determined.3.A PSO-LM combined learning algorithm is proposed to optimize the learning algorithm and convergence speed of the process neural network.Based on this algorithm,a process neural network model is established to predict the fault parameters of the photoelectric tracking subsystem.Forecasts of changes in future time horizons.And compared with the traditional algorithm,the effectiveness of the algorithm is verified.4.A radar fault prediction method based on a combination of process neural network and radial basis neural network is proposed to predict the faults of the photoelectric tracking subsystem.A design example was used to simulate the prediction process and verify the feasibility of the method.Due to the time,the research results of the dissertation have not been applied in the project.The actual results still need to be further verified by the engineering test.However,as a solution to this problem,its method can play a role in assisting decision making in radar equipment fault prediction.